Platform update93+ node packs · web, Telegram & browser automation

Automate with clarity from canvas to production

Model business logic visually, run it on a durable execution engine, and ship batch workflows plus chatbots on the web or Telegram — with browser tools, layered variables, live debugging, and secure team access.

93+
Node packs
2
Flow types
∞
Integrations
What's new

Built for teams shipping faster

Browser automation, Telegram chatbots, AI-assisted authoring, and team organization — all in one studio.

New

Browser control

Headless Chromium per flow node — navigate, click, type, and screenshot. Mint a live viewer link for human-in-the-loop login and CAPTCHA.

New

Telegram Bot

Inbound chatbot trigger with file attachments, session tracking, and optional auto-reply — same data shape as Chat Input.

New

Studio AI assistant

Describe changes in natural language; the assistant edits your canvas with tool calls grounded in your flow structure.

New

Workspaces & projects

Organize flows by team and initiative. Switch context from the home header without losing your place.

New

Context variables

Expose merged environment variables as data outputs — workspace → project → flow, with narrower scope winning on duplicates.

New

LLM agent & tool packs

Multi-provider chat models with session memory, wired to storage, vector search, databases, browsers, and MCP-style tool packs.

Flow types

One platform, two ways to run logic

The studio adapts the module library, run controls, and testing tools to match your flow type.

Workflow

Trigger-based automation with HTTP, forms, schedules, and integrations.

  • WebhookHTTP trigger with parsed body, query, and headers
  • Form RequestStart from a shareable form URL with field outputs
  • Request APICall endpoints with params, headers, and body

Chatbot

Conversational flows via web chat, Telegram, or both — with tool-augmented agents.

  • Chat InputShareable URL with message metadata and file uploads
  • Telegram BotInbound messages and attachments with optional auto-reply
  • LLM agentMulti-provider models with session memory, browser, and MCP tools

Voice Agent

SIP and VoIP call agents with speech-to-text, spoken replies, and trunk control tools.

  • Voice InputTwilio, Asterisk, FreeSWITCH, or LiveKit SIP inbound calls
  • Voice OutputTTS reply spoken back on the active call
  • SIP Trunk ToolsHangup, hold music, transfer, mute for the agent
Platform

Everything you need to ship

01

Canvas-first authoring

Compose graphs visually, duplicate nodes, and open settings panels tailored to each module — including template variables from upstream outputs.

02

Engine you can trust

Execution runs on a dedicated engine with explicit contracts per node. Install packs on demand so production only loads the integrations each flow actually uses.

03

Operational resilience

Stateful orchestration records each node's outcome. Incomplete runs resume instead of restarting from the trigger.

Node library

Built from real building blocks

Installable node packs with clear contracts, data ports, and optional tool wiring.

Chat Input

Public chat entrypoint with a shareable URL and optional file uploads — PDF, images, text, Office, and more.

Telegram Bot

Long-polling trigger for Telegram chatbots — text, attachments, session IDs, and auto-reply back to the same chat.

Browser control

Agent-wired browser tools: tabs, navigation, clicks, typing, screenshots, and optional live human viewer sessions.

LLM agent

Multi-provider chat models with session memory. Wire tools for databases, storage, vector search, and browsers in a loop.

Webhook

HTTP trigger with parsed body, query, headers, and method — start flows from any incoming request.

Schedule Trigger

Recurring schedules via a visual builder or raw cron — hourly, daily, weekly, monthly, with timezone support.

Architecture

How it works

01

Organize

Create workspaces and projects, set variables per layer, and pick Workflow, Chatbot, or Voice Agent when you add a flow.

02

Author

Drag nodes, connect ports, use the AI assistant for edits, and save definitions to your workspace.

03

Execute

The studio forwards runs to the execution engine, which processes nodes in order and persists state after each step.

04

Observe

Debug on the canvas, run chat tests for bot flows, and inspect structured outputs per step.

Ready to build your next flow?

Sign in to open the studio — workflows, chatbots, and variables are waiting in your workspace.

Sign in